snpReady: a tool to assist breeders in genomic analysis

• Published in: Molecular breeding Vol. 38; no. 8; pp. 1 - 7
• Main Authors: Granato, Italo S. C., Galli, Giovanni, de Oliveira Couto, Evellyn Giselly, e Souza, Massaine Bandeira, Mendonça, Leandro Freitas, Fritsche-Neto, Roberto
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.08.2018; Springer Nature B.V
• Subjects: Biomedical and Life Sciences; Biotechnology; Empirical analysis; Gene frequency; Genetic variability; Genomes; Genomic analysis; Genomics; Genotyping; Life Sciences; Missing data; Molecular biology; Plant biology; Plant Genetics and Genomics; Plant Pathology; Plant Physiology; Plant Sciences; Population genetics; Population number; Quality control; Genomic data; R; Software; Relationship matrix; Quality control
• ISSN: 1380-3743; 1572-9788
• DOI: 10.1007/s11032-018-0844-8

The snpReady R package is a new instrument developed to help breeders in genomic projects such as genomic prediction and association studies. This package offers three different methods to build the genomic relationship matrix, a new imputation method for missing markers based on Wright’s theory, and a population genetic overview. Therefore, we implemented three functions ( raw.data , G.matrix , and popgen ). Hence, this tool allows the raw data to be transformed from different genotyping platforms to numeric matrices and performs quality control (missing data and allele frequency). Moreover, the package generates and exports four different relationship matrices (proposed by Yang et al. (N 569:565–569, 2010), VanRaden (JDS 91:4414–23, 2008), and the Gaussian kernel) depending on the purpose and software to be used in further analysis. Finally, based on the genotypic matrix, the package estimates the genetic variability, effective population size, and endogamy, among other population genetic parameters. Empirical comparisons between the method of imputation proposed and other well-known approaches have shown a lower accuracy of imputation, however, with no significant impact on the genome prediction accuracies when a lower amount of missing data is allowed. The functions and arguments were designed to carry out the preparation of genomic datasets in a straightforward, fast, and more computationally efficient way. The package and its details are available at CRAN or http://www.github.com/italo-granato/snpReady .